The site says that the trust score was improved because the allocation for my IP is >10 years old, however my ISP didn't exist 10 years ago. The ASN for my ISP is only 5 years old (they brought some IP blocks because you know IPv4 address exhaustion!)
And very very little data is returned for my ipv6 /48
I see the same problem with rDNS. The box at the top correctly reports the DNS name as "forward-confirmed" but I see a -5 points penalty for missing reverse DNS. Tried with various IP addresses, and I observe the same for many (not all).
Oh and it doesn't even properly consume the geofeed information I have published.
BBN/successors may have held multiple class As at times, but being large ISPs probably used a lot of the space? Various clouds have a lot of space, but afaik, not in the form of whole class As.
Looks like IBM probably had multiple class As through acquisition, but I don't think they still hold them either?
https://www.hstoday.us/subject-matter-areas/cybersecurity/pe...
E.I. du Pont de Nemours (DuPont) 52.0.0.0/8
Ford Motor Company 19.0.0.0/8 Retained by Ford
Prudential Financial 48.0.0.0/8 Retained and used
like, why does Ford need a /8 to make cars? a /20 of public IPs is more than enough for corporate and virtually all of their dealerships
That data collection is about 90% solid right now, once I have it 100% I'll begin updating the frontend to read from the proper tables (currently displaying a lot of stale data). Keep checking back on the site and you'll see those changes.
This is why the website is completely free, no account or pay gates. I do plan to offer as a SaaS in time (whoever commented about batch files, I'll be offering those as well), but again this won't be done until I'm certain the captured data is 100% correct.
Thank you to those of you who gave useful feedback.
If you wanted to do something genuinely cool you could try to build an algorithmic model of latency between network points on the Internet. Such that it would be possible to estimate latency between any two network paths without sending packets first to measure it. I think I read research somewhere that this is possible and it could have applications in routing, distributed systems, and high performance networking.